Internal testing discovered security risks, and it was reported that OpenAI canceled the release of AI model GPT‑6.1 Astra内部测试发现安全隐患，消息称 OpenAI 取消发布 AI 模型 GPT‑6.1 Astra

ITHome reported on September 29 that according to the Wall Street Journal, OpenAI decided to cancel the release of GPT‑6.1 Astra due to multiple security risks raised by researchers during internal testing. This next-generation AI model was originally scheduled to be officially unveiled in October. According to reports, the model was originally scheduled to be deployed in ChatGPT and Codex products, and was designed to handle more complex tasks without human assistance. Earlier this month, Anthropic CEO Dario Amodei called on the industry to slow down the development of cutting-edge AI models so that security measures can keep up with technology iterations. OpenAI CEO Sam Altman and SpaceX CEO Elon Musk both agree. Saachi Jain, OpenAI's head of security, told the Wall Street Journal on Monday that Astra failed to meet the company's internal standards during alignment testing, which is u…